CN112307202A - Text information correction method, cloud computing system and computer storage medium - Google Patents
Text information correction method, cloud computing system and computer storage medium Download PDFInfo
- Publication number
- CN112307202A CN112307202A CN201910690944.9A CN201910690944A CN112307202A CN 112307202 A CN112307202 A CN 112307202A CN 201910690944 A CN201910690944 A CN 201910690944A CN 112307202 A CN112307202 A CN 112307202A
- Authority
- CN
- China
- Prior art keywords
- text information
- neural network
- trained neural
- network model
- cloud computing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/10—File systems; File servers
- G06F16/18—File system types
- G06F16/1805—Append-only file systems, e.g. using logs or journals to store data
- G06F16/1815—Journaling file systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Databases & Information Systems (AREA)
- Machine Translation (AREA)
Abstract
The embodiment of the invention discloses a method for correcting text information, which comprises the following steps: when the cloud computing system has a fault, acquiring text information of the fault error report from a log file of the cloud computing system, when the text information represents that the configuration item of the component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information, and correcting the configuration item of the component corresponding to the text information according to the correction information to enable the cloud computing system to normally operate. The embodiment of the invention also discloses a cloud computing system and a computer storage medium, which improve the efficiency of fault repair of the cloud computing system and are beneficial to the normal operation of the cloud computing system.
Description
Technical Field
The invention relates to a repair technology of a cloud computing system, in particular to a text information correction method, a cloud computing system and a computer storage medium.
Background
In the field of cloud computing, OpenStack is an open-source cloud computing management platform project, and is combined by several main components to complete specific work. OpenStack has obtained wide application range and market share, has become a practical standard in the private cloud field, has obtained huge influence in the public cloud field, and many enterprises and public institutions are all using OpenStack to establish the cloud computing platform of host computer, and OpenStack is developing rapidly at home and abroad. However, OpenStack contains many components, and besides the main components of computation, storage and network, there are many other components, and with the development of OpenStack, there are still more components added.
In the OpenStack deployment and operation and maintenance process, a fault caused by configuration item configuration errors occupies a very large proportion, however, since OpenStack has a very large number of components and configuration items of each component are complicated, in the development and operation and maintenance process, manual troubleshooting is adopted, so that too much time is occupied, and the efficiency is very low; therefore, the technical problem of low efficiency exists in the mode that the configuration errors of the configuration items are manually checked and repaired in the conventional cloud computing system.
Disclosure of Invention
In view of this, embodiments of the present invention are expected to provide a method for correcting text information, a cloud computing system, and a computer storage medium, so as to solve the technical problem that in the prior art, a configuration error of a cloud computing system is repaired by manual troubleshooting, and the efficiency is low.
The technical scheme of the application is realized as follows:
in a first aspect, an embodiment of the present application provides a method for correcting text information, where the method includes:
when a cloud computing system fails, acquiring text information of failure error report from a log file of the cloud computing system;
when the text information represents that the configuration item of the component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information;
and correcting the configuration items of the components corresponding to the text information according to the correction information so as to enable the cloud computing system to normally operate.
In the above method, the method further comprises:
acquiring historical text information of error reporting of configuration items of the components from a log file of the cloud computing system; the number of the historical text information is a preset number;
selecting N groups of text information from the historical text information; each group of text information at least comprises two pieces of text information;
vectorizing the N groups of text information respectively to obtain N groups of processed text information;
and performing optimization training on a preset neural network model according to the N groups of processed text information to obtain the trained neural network model.
In the above method, the selecting N groups of text information from the historical text information includes:
determining an error report category of the historical text information by using a preset error report keyword, and classifying the historical text information according to the error report category of the historical text information to obtain a first-class subset;
classifying the first type subset according to the component type of the cloud computing system to obtain a second type subset;
and respectively selecting text information from each subset in the second type of subsets, and forming a group of text information by using the selected text information until N groups of text information are formed.
In the above method, the performing optimization training on a preset neural network model according to the N groups of processed text information to obtain the trained neural network model includes:
repeatedly selecting N times from the N groups of processed text information, and selecting N training sample sets; each training sample set comprises N-1 groups of text information, and any two training sample sets are different;
acquiring an ith training sample set; wherein i is the number of times of optimization training;
when i is equal to 1, performing optimization training on a preset neural network model by adopting an ith training sample set to obtain an ith trained neural network model;
and when i is larger than 1 and is smaller than or equal to N, performing optimization training on the i-1 th trained neural network model by adopting the ith training sample set to obtain the ith trained neural network model, updating i to i +1 until the Nth trained neural network model is obtained, and taking the Nth trained neural network model as the trained neural network model.
In the above method, after repeating the selection N times from N groups of processed text information and selecting N training sample sets, the method further includes:
respectively adopting N training sample sets to carry out optimization training on a preset neural network model to obtain N trained neural network models;
and determining the trained neural network model according to the N trained neural network models.
In the above method, the determining the trained neural network model according to the N trained neural network models includes:
selecting text information from each subset in the second type subsets after N groups of text information are removed, and forming a group of text information by using the selected text information until K groups of text information are formed;
according to the K groups of text information, respectively carrying out accuracy rate tests on the N trained neural network models to determine the accuracy rates of the N trained neural network models;
and selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and determining the trained neural network model as the trained neural network model.
In the above method, after selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models in order of high accuracy to low accuracy and determining the selected trained neural network model as the trained neural network model, the method further includes:
selecting the first M trained neural network models with the accuracy ranked from the N trained neural network models according to the sequence of the accuracy from high to low; wherein M is less than N;
acquiring training sample sets corresponding to the first M trained neural network models;
grouping the processed text information in the training sample sets corresponding to the first M trained neural network models again to obtain j groups of training sample sets;
and performing optimization training again on the trained neural network model by adopting j groups of training sample sets, and updating the trained neural network model by using the neural network model obtained by performing optimization training again.
In a second aspect, an embodiment of the present invention provides a cloud computing system, where the cloud computing system includes:
the device comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring text information of fault error report from a log file of a cloud computing system when the cloud computing system has a fault;
the processing unit is used for processing the text information by adopting a trained neural network model when the text information represents that the configuration item of the component of the cloud computing system is wrong, so as to obtain the correction information of the configuration item of the component corresponding to the text information;
and the correction unit is used for correcting the configuration item of the component corresponding to the text information according to the correction information so as to enable the cloud computing system to normally operate.
In the above cloud computing system, the cloud computing system further includes: a training unit, the training unit comprising:
the acquisition subunit is used for acquiring historical text information of configuration item errors of the components from a log file of the cloud computing system;
the number of the historical text information is a preset number;
the selecting subunit is used for selecting N groups of text information from the historical text information;
each group of text information at least comprises two pieces of text information;
the processing subunit is used for respectively carrying out vectorization processing on the N groups of text information to obtain N groups of processed text information;
and the training subunit is used for carrying out optimization training on a preset neural network model according to the N groups of processed text information to obtain the trained neural network model.
In the cloud computing system, the selecting subunit is specifically configured to:
determining an error report category of the historical text information by using a preset error report keyword, and classifying the historical text information according to the error report category of the historical text information to obtain a first-class subset;
classifying the first type subset according to the component type of the cloud computing system to obtain a second type subset;
and respectively selecting text information from each subset in the second type of subsets, and forming a group of text information by using the selected text information until N groups of text information are formed.
In the cloud computing system, the training subunit is specifically configured to:
repeatedly selecting N times from the N groups of processed text information, and selecting N training sample sets;
each training sample set comprises N-1 groups of text information, and any two training sample sets are different;
acquiring an ith training sample set; wherein i is the number of times of optimization training;
when i is equal to 1, performing optimization training on a preset neural network model by adopting an ith training sample set to obtain an ith trained neural network model;
and when i is larger than 1 and is smaller than or equal to N, performing optimization training on the i-1 th trained neural network model by adopting the ith training sample set to obtain the ith trained neural network model, updating i to i +1 until the Nth trained neural network model is obtained, and taking the Nth trained neural network model as the trained neural network model.
In the cloud computing system, the training unit further includes:
the reselecting subunit is used for repeatedly selecting N times from the N groups of processed text information, and after N training sample sets are selected, optimizing and training the preset neural network model by respectively adopting the N training sample sets to obtain N trained neural network models;
and the first determining subunit is used for determining the trained neural network model according to the N trained neural network models.
In the cloud computing system, the first determining subunit is specifically configured to:
selecting text information from each subset in the second type subsets after N groups of text information are removed, and forming a group of text information by using the selected text information until K groups of text information are formed;
according to the K groups of text information, respectively carrying out accuracy rate tests on the N trained neural network models to determine the accuracy rates of the N trained neural network models;
and selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and determining the trained neural network model as the trained neural network model.
In the cloud computing system, the training unit further includes:
the second determining subunit is used for selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and selecting the trained neural network models with the accuracy ranked in the first M from the N trained neural network models according to the sequence of the accuracy from high to low after determining the trained neural network models; wherein M is less than N;
acquiring training sample sets corresponding to the first M trained neural network models;
grouping the processed text information in the training sample sets corresponding to the first M trained neural network models again to obtain j groups of training sample sets;
and performing optimization training again on the trained neural network model by adopting j groups of training sample sets, and updating the trained neural network model by using the neural network model obtained by performing optimization training again.
In a third aspect, an embodiment of the present application further provides a cloud computing system, where the cloud computing system includes: the device comprises a processor and a storage medium storing instructions executable by the processor, wherein the storage medium depends on the processor to execute operations through a communication bus, and when the instructions are executed by the processor, the method for correcting the text information is executed.
The embodiment of the application provides a computer storage medium, which stores executable instructions, and when the executable instructions are executed by one or more processors, the processors execute the text information correction method of one or more embodiments.
The embodiment of the invention provides a text information correction method, a cloud computing system and a computer storage medium, wherein the method comprises the following steps: firstly, when a cloud computing system has a fault, acquiring text information of fault error report from a log file of the cloud computing system, when the text information represents that a configuration item of a component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information, and correcting the configuration item of the component corresponding to the text information according to the correction information to enable the cloud computing system to work normally; that is to say, in the embodiment of the present invention, the text information of the fault is acquired from the log file of the cloud computing system, and when the text information indicates that the configuration item of the component of the cloud computing system is wrong, the trained neural network model is used to process the text information to obtain the correction information of the component configuration item corresponding to the text information, so that the fault of the wrong configuration item can be automatically repaired, and the fault repair is avoided being performed in a manual troubleshooting manner, thereby improving the fault repair efficiency, and facilitating the normal operation of the cloud computing system.
Drawings
Fig. 1 is a schematic flow chart of an optional text message correction method in an embodiment of the present invention;
FIG. 2A is a first schematic structural diagram of a deep learning model;
FIG. 2B is a schematic structural diagram of a deep learning model II;
FIG. 3 is a schematic diagram of a convolutional neural network;
fig. 4 is a flowchart illustrating an example of another optional text information modification method in the embodiment of the present invention;
FIG. 5 is a first schematic structural diagram of a cloud computing system according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a cloud computing system in the embodiment of the present invention.
Detailed Description
The technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.
An embodiment of the present invention provides a method for correcting text information, where fig. 1 is a schematic flow diagram of a method for correcting optional text information in an embodiment of the present invention, and as shown in fig. 1, the method for correcting text information may include:
s101: when the cloud computing system fails, acquiring text information of failure error report from a log file of the cloud computing system;
at present, cloud computing has wide application due to its advantages of large scale, virtualization, high reliability, strong versatility, expandability, and low price, for example, for the private cloud field, if a private cloud is to be constructed or a service or development is provided by using its own hardware and computing capability, a cloud computing platform of its own must be built, the cloud computing platforms that are open on the network mainly include Abicloud, OpenStack, and the like, and the following description will be given by taking OpenStack as an example of a cloud computing system.
In practical application, for an OpenStack-based cloud computing system, in a deployment or operation and maintenance process, if a cloud computing system has a fault with a wrong configuration item, a manual troubleshooting mode is usually adopted, but the manual troubleshooting mode is time-consuming and labor-consuming, and operation and maintenance efficiency of the cloud computing system is affected.
In order to improve the operation and maintenance efficiency of the cloud computing system, firstly, when a fault of the cloud computing system is detected, a log file is obtained from the cloud computing system, the log file is a recording file or a file set used for recording the operation time of the cloud computing system, and after the obtained log file of the cloud computing system, text information of fault error report can be obtained from the log file.
S102: when the text information represents that the configuration item of the component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information;
after the text information is obtained through the S101, the type of the fault of the text information is judged, and after the judgment, when the text information represents that a certain hardware in the cloud computing system has a fault and reports the fault, the cloud computing system generates a prompt message for prompting the certain hardware in the computing system to have the fault.
When the text information represents that the configuration item of the component of the cloud computing system is wrong, that is, the reason that the cloud computing system fails is that the configuration item of the component in the cloud computing system is configured incorrectly, in order to correct the error in the configuration item, the text information may be trained by using a pre-trained neural network to obtain correction information of the configuration item of the component corresponding to the text information.
The trained neural network model is obtained by optimizing and training a preset neural network model through a training sample set in advance.
In practical application, a convolutional neural network in a deep learning framework is usually adopted as a neural network model, the deep learning model is an artificial neural network comprising a plurality of hidden layers, each layer in the network is used as the input of the next layer, and then the actual output result of the network to the original input information is made to be as close to the target output result as possible by continuously adjusting network parameters; the artificial neural network has more hidden layers, and the parameters in the artificial neural network are initialized layer by layer, which are two main characteristics of a deep learning model.
FIG. 2A is a schematic structural diagram of a deep learning model, wherein FIG. 2A is a shallow deep learning model with a hidden layer, as shown in FIG. 2A, the shallow deep learning model includes an input layer, a hidden layer and an output layer; fig. 2B is a structural diagram of a deep learning model, wherein fig. 2B is a deep learning model including a plurality of hidden layers, and as shown in fig. 2B, the deep learning model includes an input layer, a plurality of hidden layers, and an output layer.
In practical application, the convolutional neural network is a variation of a deep learning model, is a multilayer feedforward neural network, is derived from the research result of Wiesel et al on the cerebral visual cortex of cats, is inspired by the convolutional neural network, and adopts three methods for realizing abstraction and simulation of the cerebral nervous system of primates: local receptive field, weight sharing and down sampling.
The convolutional neural network is used as a deep learning model, has the advantages of being capable of automatically learning and extracting features from input original data directly, and is obviously superior to a traditional method in generalization capability. A convolutional neural network is a neural network comprising a plurality of layers, each layer having a plurality of two-dimensional planes, each two-dimensional plane comprising a plurality of neurons.
Fig. 3 is a schematic structural diagram of a convolutional neural network, and as shown in fig. 3, the convolutional neural network includes: an input layer, a hidden layer (divided into a convolutional layer and a downsampling layer), and an output layer; the convolutional layer and the downsampling layer form a feature extraction stage, usually, the convolutional neural network has 1-3 feature extraction stages, and fig. 3 has two feature extraction stages, and the plurality of feature extraction stages are formed by alternately arranging and stacking the plurality of convolutional layers and the downsampling layers; after the input original information is subjected to feature extraction of a plurality of convolutional layers and downsampling layers, high-level features with strong expression capability are obtained at the rear end of a convolutional neural network, and finally the features are output to a trainable classifier to perform classification and identification tasks.
Here, it should be noted that the trained neural network model may be trained based on the convolutional neural network, or may be trained based on a Back Propagation (BP) neural network model, and the embodiment of the present invention is not limited in particular.
In order to obtain the trained neural network model, in an alternative embodiment, the method for correcting the text information further includes:
s401: acquiring historical text information of configuration item errors of the components from a log file of the cloud computing system;
the number of the historical text information is a preset number;
s402: selecting N groups of text information from the historical text information;
each group of text information at least comprises two pieces of text information;
firstly, in order to obtain a trained neural network model, a training sample set needs to be obtained first, historical text information of configuration items of components which are reported by mistake can be obtained from a log file of a cloud computing system, the training sample set is selected from the historical text information, and in order to achieve a better training effect, the preset number can be more than thousands or more than ten thousands, for example, 5000 pieces of text information.
It should be noted that the more the number and the more the types of the collected historical text information are, the stronger the correction capability of the trained neural network model on the text information is; therefore, when acquiring the history text information, as many kinds of history text information as possible are acquired.
Then, N groups of text information are selected from the historical text information as a training sample set for training a preset neural network model, and in order to make the trained neural network model have a strong ability to correct the text information, there is a certain requirement on the number and the type of the text information in the training sample set, in an alternative embodiment, S402 may include:
determining an error report category of the historical text information by using a preset error report keyword, and classifying the historical text information according to the error report category of the historical text information to obtain a first-class subset;
classifying the first type subset according to the component type of the cloud computing system to obtain a second type subset;
and respectively selecting text information from each subset in the second type of subsets, and forming a group of text information by using the selected text information until N groups of text information are formed.
Specifically, information extraction is performed on historical text information according to an error-reporting keyword, for example, the extracted error-reporting keyword may be "NameError", and information with the error-reporting keyword being "NameError" is extracted, for example, the text information is "memanic driver openvswitch failed in bound _ port, the NameError is that" global name 'port' is not defined ", and thus, the error-reporting category of the text information can be determined, and" NameError "of the same category is divided into a category subset, so that a first category subset is obtained.
Then, according to component categories of the cloud computing system, for example, a storage service component, a shared server component, a background service component, an add operation component, a delete operation component, a modify operation component, and a find operation component in OpenStack, for example, components of an Application Programming Interface (API) of an Application program related to a network (network) may include:
it can be seen that the components of OpenStack are numerous, such as add-drop-and-modify-and-check operations of the respective components.
For example, for each subset of the same error reporting category in the first-class subset, the text information that is the same as the stored service components is divided into a subset, the text information that is the same as the shared service components is divided into a subset, and so on, so as to obtain several subsets of different error reporting categories and different component categories, and these subsets are used as the second-class subset.
In this way, each subset in the second type of subsets represents text information of one error report type and one component type, and then the text information is selected from each subset in the second type of subsets respectively, and a group of text information is formed by using the selected text information, so that the formed group of text information comprises text information of multiple types, and N groups of text information are formed, so that each group of text information comprises text information of multiple types, and thus, enough number of text information is obtained, the number of N groups of text information is as large as possible, the types are complete, and the correction effect of the trained neural network model is better.
S403: vectorizing the N groups of text information respectively to obtain N groups of processed text information;
s404: and performing optimization training on the preset neural network model according to the N groups of processed text information to obtain a trained neural network model.
Specifically, after N groups of text information are obtained, vectorization processing needs to be performed on the text information to obtain N groups of processed text information, and the N groups of processed text information are used to train a preset neural network model, and in practical applications, the vectorization processing may use a Word2Vec technology or a Doc2Vec technology, where embodiments of the present invention are not specifically limited.
To obtain a trained neural network model, in an alternative embodiment, S404 may include:
repeatedly selecting N times from the N groups of processed text information, and selecting N training sample sets;
each training sample set comprises N-1 groups of text information, and any two training sample sets are different;
acquiring an ith training sample set; wherein i is the number of times of optimization training;
when i is equal to 1, performing optimization training on a preset neural network model by adopting an ith training sample set to obtain an ith trained neural network model;
and when i is larger than 1 and is smaller than or equal to N, performing optimization training on the i-1 th trained neural network model by adopting the ith training sample set to obtain the ith trained neural network model, updating i to i +1 until the Nth trained neural network model is obtained, and taking the Nth trained neural network model as the trained neural network model.
Here, the training sample set is determined based on N groups of processed text information, specifically, N-1 groups of text information are selected from the N groups of processed text information, the selection is repeated N times, N-1 groups of processed text information can be selected, and each N-1 group of processed text information is determined as one training sample set, so as to obtain N training sample sets.
And then, carrying out iterative optimization training, determining the number of iterations to be N, training a preset neural network model by using a 1 st training sample set during 1 st iteration to obtain a 1 st trained neural network model, training the 1 st trained neural network model by using a 2 nd training sample set during 2 nd iteration to obtain a 2 nd trained neural network model, and repeating the steps to obtain the Nth trained neural network model.
Therefore, through an iterative optimization training mode, the correction capability of the trained neural network model is stronger when the iteration times are more.
In order to obtain a trained neural network model, in an alternative embodiment, after repeatedly selecting N times from N sets of processed text information and selecting N training sample sets, the method further includes:
respectively adopting N training sample sets to carry out optimization training on a preset neural network model to obtain N trained neural network models;
and determining the trained neural network model according to the N trained neural network models.
That is to say, after N training sample sets are determined, N training sample sets may be used to train the preset neural network models, respectively, so as to obtain N trained neural network models, and finally, the trained neural network models are determined according to the N trained neural network models.
In order to obtain the trained neural network model, in an alternative embodiment, determining the trained neural network model according to N trained neural network models includes:
selecting text information from each subset in the second type subsets after N groups of text information are removed, and forming a group of text information by using the selected text information until K groups of text information are formed;
according to the K groups of text information, respectively carrying out accuracy rate tests on the N trained neural network models to determine the accuracy rates of the N trained neural network models;
and selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and determining the trained neural network model as the trained neural network model.
That is, based on the second subset after the N groups of text information are selected, the text information is selected from each subset, a group of text information is formed by using the selected text information until K groups of text information are formed, the K groups of text information are used as a test training sample set, accuracy tests are respectively performed on the N trained neural network models, the accuracy of the N trained neural network models is determined according to the obtained N trained neural network model correction information, the N trained neural network models are ranked according to the accuracy, the trained neural network model corresponding to the highest accuracy is selected from the N trained neural network models, and the model is considered as the optimal model, so that the model is finally determined as the trained neural network model.
In order to obtain an optimal trained neural network model, in an optional embodiment, after selecting the trained neural network model corresponding to the highest accuracy value from the N trained neural network models in the order from high accuracy to low accuracy and determining the trained neural network model, the method further includes:
selecting the first M trained neural network models with the accuracy ranked from the N trained neural network models according to the sequence of the accuracy from high to low;
wherein M is less than N;
acquiring training sample sets corresponding to the first M trained neural network models;
grouping the processed text information in the training sample sets corresponding to the first M trained neural network models again to obtain j groups of training sample sets;
and performing optimization training again on the trained neural network model by adopting j groups of training sample sets, and determining the neural network model obtained by performing optimization training again as the trained neural network model.
Specifically, after N trained neural network models are determined, the training sample set can be re-determined according to the accuracy, the trained neural network models are continuously trained, and the neural network models obtained through re-optimization training are determined as the trained neural network models.
S103: and correcting the configuration items of the components corresponding to the text information according to the correction information so as to enable the cloud computing system to normally operate.
After obtaining the correction information, knowing which component configuration item has an error, it may determine the correct information of the configuration item, where the correction information includes the type of the configuration item, the storage location, the correct information of the configuration item, and the like, and this is not limited in this embodiment of the present invention.
Accessing the optimized network model obtained by training into an OpenStack environment, monitoring log information of each component in real time, automatically starting the automatic repair function through RPC communication when error reporting information related to suspected configuration errors is found, automatically identifying corresponding error configuration item information through the corresponding error reporting log information by the network model, and performing corresponding correction, and if the error is not a fault caused by configuration item errors, notifying corresponding development and operation and maintenance personnel to perform problem troubleshooting in other aspects through prompt information.
According to the method in one or more embodiments, in the OpenStack automatic repair process, a convolutional neural network in a deep learning framework is adopted, when an optimized network model is trained and learned, an iterative training learning method is adopted, the neural network model is accessed to the OpenStack environment, log files of all components are monitored in real time, and when error reporting information related to suspected configuration errors is found, the automatic repair function is started automatically through Remote Procedure Call (RPC) communication.
The embodiment of the invention provides a method for correcting text information, which comprises the following steps: firstly, when a cloud computing system has a fault, acquiring text information of fault error report from a log file of the cloud computing system, when the text information represents that a configuration item of a component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information, and correcting the configuration item of the component corresponding to the text information according to the correction information to enable the cloud computing system to work normally; that is to say, in the embodiment of the present invention, the text information of the fault is acquired from the log file of the cloud computing system, and when the text information indicates that the configuration item of the component of the cloud computing system is wrong, the trained neural network model is used to process the text information to obtain the correction information of the component configuration item corresponding to the text information, so that the fault of the configuration item error can be automatically repaired, and the fault troubleshooting is avoided in a manner of using a user to perform troubleshooting, so that the fault repairing efficiency is improved, and the normal operation of the cloud computing system is facilitated.
Based on the same inventive concept, this embodiment provides a cloud computing system, fig. 5 is a schematic structural diagram of the cloud computing system in the embodiment of the present invention, as shown in fig. 5, the cloud computing system includes: an acquisition unit 51, a processing unit 52, and a correction unit 53;
the acquiring unit 51 is used for acquiring text information of fault error report from a log file of the cloud computing system when the cloud computing system has a fault;
the processing unit 52 is configured to, when the text information represents that the configuration item of the component of the cloud computing system is wrong, process the text information by using the trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information;
and the correcting unit 53 is configured to correct the configuration item of the component corresponding to the text message according to the correction information, so that the cloud computing system operates normally.
In an optional embodiment, the cloud computing system further comprises: a training unit, the training unit comprising:
the acquisition subunit is used for acquiring historical text information of configuration item errors of the components from a log file of the cloud computing system;
the number of the historical text information is a preset number;
the selecting subunit is used for selecting N groups of text information from the historical text information;
each group of text information at least comprises two pieces of text information;
the processing subunit is used for respectively carrying out vectorization processing on the N groups of text information to obtain N groups of processed text information;
and the training subunit is used for carrying out optimization training on the preset neural network model according to the N groups of processed text information to obtain the trained neural network model.
In an optional embodiment, the selecting subunit is specifically configured to:
determining an error report type of the historical text information by using a preset error report keyword, and classifying the historical text information according to the error report type of the historical text information to obtain a first type subset;
classifying the first type subset according to the component type of the cloud computing system to obtain a second type subset;
and respectively selecting text information from each subset in the second type of subsets, and forming a group of text information by using the selected text information until N groups of text information are formed.
In an alternative embodiment, the training subunit is specifically configured to:
repeatedly selecting N times from the N groups of processed text information, and selecting N training sample sets;
each training sample set comprises N-1 groups of text information, and any two training sample sets are different;
acquiring an ith training sample set; wherein i is the number of times of optimization training;
when i is equal to 1, performing optimization training on a preset neural network model by adopting an ith training sample set to obtain an ith trained neural network model;
and when i is larger than 1 and is smaller than or equal to N, performing optimization training on the i-1 th trained neural network model by adopting the ith training sample set to obtain the ith trained neural network model, updating i to i +1 until the Nth trained neural network model is obtained, and taking the Nth trained neural network model as the trained neural network model.
In an alternative embodiment, the training unit further comprises:
the reselecting subunit is used for repeatedly selecting N times from the N groups of processed text information, and after N training sample sets are selected, optimizing and training the preset neural network model by respectively adopting the N training sample sets to obtain N trained neural network models;
and the first determining subunit is used for determining the trained neural network model according to the N trained neural network models.
In an optional embodiment, the first determining subunit is specifically configured to:
selecting text information from each subset in the second type subsets after N groups of text information are removed, and forming a group of text information by using the selected text information until K groups of text information are formed;
according to the K groups of text information, respectively carrying out accuracy rate tests on the N trained neural network models to determine the accuracy rates of the N trained neural network models;
and selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and determining the trained neural network model as the trained neural network model.
In an optional embodiment, the training unit further comprises:
the second determining subunit is used for selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and selecting the trained neural network models with the accuracy ranked in the first M from the N trained neural network models according to the sequence of the accuracy from high to low after determining the trained neural network models;
wherein M is less than N;
acquiring training sample sets corresponding to the first M trained neural network models;
grouping the processed text information in the training sample sets corresponding to the first M trained neural network models again to obtain j groups of training sample sets;
and performing optimization training again on the trained neural network model by adopting j groups of training sample sets, and updating the trained neural network model by using the neural network model obtained by performing optimization training again.
In practical applications, the obtaining Unit 51, the Processing Unit 52, the modifying Unit 53, the obtaining sub-Unit, the selecting sub-Unit, the Processing sub-Unit, the training sub-Unit, the re-selecting sub-Unit, the first determining sub-Unit, and the second determining sub-Unit may be implemented by a processor located on a cloud computing system, specifically, implemented by a Central Processing Unit (CPU), a Microprocessor Unit (MPU), a Digital Signal Processor (DSP), or a Field Programmable Gate Array (FPGA).
Fig. 6 is a schematic structural diagram of a cloud computing system provided in an embodiment of the present application, and as shown in fig. 6, an embodiment of the present application provides a cloud computing system 600, including:
a processor 61 and a storage medium 62 storing instructions executable by the processor 61, wherein the storage medium 62 depends on the processor 61 to perform operations through a communication bus 63, and when the instructions are executed by the processor 61, the method for correcting the text message according to the first embodiment is performed.
It should be noted that, in practical applications, the various components in the terminal are coupled together by a communication bus 63. It will be appreciated that the communication bus 63 is used to enable communications among the components. The communication bus 63 includes a power bus, a control bus, and a status signal bus, in addition to a data bus. But for clarity of illustration the various buses are labeled in figure 6 as communication bus 63.
The embodiment of the application provides a computer storage medium, which stores executable instructions, and when the executable instructions are executed by one or more processors, the processors execute the method for correcting text information according to the first embodiment.
The computer-readable storage medium may be a magnetic random access Memory (FRAM), a Read Only Memory (ROM), a Programmable Read Only Memory (PROM), an Erasable Programmable Read Only Memory (EPROM), an Electrically Erasable Programmable Read Only Memory (EEPROM), a Flash Memory (Flash Memory), a magnetic surface Memory, an optical Disc, or a Compact Disc Read-Only Memory (CD-ROM), among others.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present application, and is not intended to limit the scope of the present application.
Claims (10)
1. A method for correcting text information, comprising:
when a cloud computing system fails, acquiring text information of failure error report from a log file of the cloud computing system;
when the text information represents that the configuration item of the component of the cloud computing system is wrong, processing the text information by adopting a trained neural network model to obtain correction information of the configuration item of the component corresponding to the text information;
and correcting the configuration items of the components corresponding to the text information according to the correction information so as to enable the cloud computing system to normally operate.
2. The method of claim 1, further comprising:
acquiring historical text information of error reporting of configuration items of the components from a log file of the cloud computing system; the number of the historical text information is a preset number;
selecting N groups of text information from the historical text information; each group of text information at least comprises two pieces of text information;
vectorizing the N groups of text information respectively to obtain N groups of processed text information;
and performing optimization training on a preset neural network model according to the N groups of processed text information to obtain the trained neural network model.
3. The method of claim 2, wherein the selecting N groups of text information from the historical text information comprises:
determining an error report category of the historical text information by using a preset error report keyword, and classifying the historical text information according to the error report category of the historical text information to obtain a first-class subset;
classifying the first type subset according to the component type of the cloud computing system to obtain a second type subset;
and respectively selecting text information from each subset in the second type of subsets, and forming a group of text information by using the selected text information until N groups of text information are formed.
4. The method according to claim 2 or 3, wherein the performing optimization training on a preset neural network model according to the N groups of processed text information to obtain the trained neural network model comprises:
repeatedly selecting N times from the N groups of processed text information, and selecting N training sample sets; each training sample set comprises N-1 groups of text information, and any two training sample sets are different;
acquiring an ith training sample set; wherein i is the number of times of optimization training;
when i is equal to 1, performing optimization training on a preset neural network model by adopting an ith training sample set to obtain an ith trained neural network model;
and when i is larger than 1 and is smaller than or equal to N, performing optimization training on the i-1 th trained neural network model by adopting the ith training sample set to obtain the ith trained neural network model, updating i to i +1 until the Nth trained neural network model is obtained, and taking the Nth trained neural network model as the trained neural network model.
5. The method of claim 4, wherein after repeating the selecting N times from the N sets of processed text information, and selecting N training sample sets, the method further comprises:
respectively adopting N training sample sets to carry out optimization training on a preset neural network model to obtain N trained neural network models;
and determining the trained neural network model according to the N trained neural network models.
6. The method of claim 5, wherein determining the trained neural network model from the N trained neural network models comprises:
selecting text information from each subset in the second type subsets after N groups of text information are removed, and forming a group of text information by using the selected text information until K groups of text information are formed;
according to the K groups of text information, respectively carrying out accuracy rate tests on the N trained neural network models to determine the accuracy rates of the N trained neural network models;
and selecting the trained neural network model corresponding to the highest accuracy from the N trained neural network models according to the sequence of the accuracy from high to low, and determining the trained neural network model as the trained neural network model.
7. The method of claim 6, wherein after the trained neural network model corresponding to the highest accuracy value is selected from the N trained neural network models in order of high accuracy to low accuracy and determined as the trained neural network model, the method further comprises:
selecting the first M trained neural network models with the accuracy ranked from the N trained neural network models according to the sequence of the accuracy from high to low; wherein M is less than N;
acquiring training sample sets corresponding to the first M trained neural network models;
grouping the processed text information in the training sample sets corresponding to the first M trained neural network models again to obtain j groups of training sample sets;
and performing optimization training again on the trained neural network model by adopting j groups of training sample sets, and updating the trained neural network model by using the neural network model obtained by performing optimization training again.
8. A cloud computing system, the cloud computing system comprising:
the device comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is used for acquiring text information of fault error report from a log file of a cloud computing system when the cloud computing system has a fault;
the processing unit is used for processing the text information by adopting a trained neural network model when the text information represents that the configuration item of the component of the cloud computing system is wrong, so as to obtain the correction information of the configuration item of the component corresponding to the text information;
and the correction unit is used for correcting the configuration item of the component corresponding to the text information according to the correction information so as to enable the cloud computing system to normally operate.
9. A cloud computing system, the cloud computing system comprising:
a processor and a storage medium storing instructions executable by the processor, the storage medium performing operations by relying on the processor through a communication bus, the instructions when executed by the processor performing the method of correcting text information according to any one of claims 1 to 7.
10. A computer storage medium having stored thereon executable instructions that, when executed by one or more processors, perform the method of modifying textual information of any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910690944.9A CN112307202B (en) | 2019-07-29 | 2019-07-29 | Text information correction method, cloud computing system and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910690944.9A CN112307202B (en) | 2019-07-29 | 2019-07-29 | Text information correction method, cloud computing system and computer storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112307202A true CN112307202A (en) | 2021-02-02 |
| CN112307202B CN112307202B (en) | 2022-12-13 |
Family
ID=74328943
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910690944.9A Active CN112307202B (en) | 2019-07-29 | 2019-07-29 | Text information correction method, cloud computing system and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112307202B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107707392A (en) * | 2017-09-26 | 2018-02-16 | 厦门集微科技有限公司 | Passage restorative procedure and device, terminal |
| CN109063785A (en) * | 2018-08-23 | 2018-12-21 | 国网河北省电力有限公司沧州供电分公司 | charging pile fault detection method and terminal device |
| CN109445935A (en) * | 2018-10-10 | 2019-03-08 | 杭州电子科技大学 | A kind of high-performance big data analysis system self-adaption configuration method under cloud computing environment |
-
2019
- 2019-07-29 CN CN201910690944.9A patent/CN112307202B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107707392A (en) * | 2017-09-26 | 2018-02-16 | 厦门集微科技有限公司 | Passage restorative procedure and device, terminal |
| CN109063785A (en) * | 2018-08-23 | 2018-12-21 | 国网河北省电力有限公司沧州供电分公司 | charging pile fault detection method and terminal device |
| CN109445935A (en) * | 2018-10-10 | 2019-03-08 | 杭州电子科技大学 | A kind of high-performance big data analysis system self-adaption configuration method under cloud computing environment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112307202B (en) | 2022-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110366734B (en) | Optimizing neural network architecture | |
| EP3583553A1 (en) | Neural architecture search for convolutional neural networks | |
| CN108563555B (en) | Fault change code prediction method based on four-target optimization | |
| US20210141779A1 (en) | System and method for facilitating an objective-oriented data structure and an objective via the data structure | |
| CN109656802A (en) | Test case design system based on high coupling automatic matching technology | |
| CN109886554A (en) | Unlawful practice method of discrimination, device, computer equipment and storage medium | |
| CN113761525A (en) | Intelligent intrusion detection method and system based on federal learning | |
| CN114328277A (en) | Software defect prediction and quality analysis method, device, equipment and medium | |
| CN114237754B (en) | Data loading method and device, electronic equipment and storage medium | |
| CN108604313B (en) | Automated predictive modeling and framework | |
| CN116340726A (en) | Energy economy big data cleaning method, system, equipment and storage medium | |
| CN103440460A (en) | Application system change validation method and system | |
| CN112154415A (en) | Efficient event management in a mainframe computer system | |
| US11620550B2 (en) | Automated data table discovery for automated machine learning | |
| US20220012643A1 (en) | Account prediction using machine learning | |
| CN112307202B (en) | Text information correction method, cloud computing system and computer storage medium | |
| CN117033209A (en) | AI model training method, BIOS testing method, device, equipment and storage medium | |
| CN117591961A (en) | Pulsar candidate identification method and pulsar candidate identification system based on self-normalization classification model | |
| CN114139636B (en) | Abnormal operation processing method and device | |
| CN111950753A (en) | Scenic spot passenger flow prediction method and device | |
| CN117236304A (en) | Method for realizing Excel general import based on template configuration | |
| CN110716778A (en) | Application compatibility testing method, device and system | |
| CN116402325A (en) | Automatic business process processing method and device | |
| CN106709572A (en) | Data processing method and equipment | |
| CN103942403A (en) | Method and device for screening mass variables |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |